EP0370422A2 - Verfahren und Gerät zur Fehlermessung und -verringerung für Massenspeichersystem - Google Patents

Verfahren und Gerät zur Fehlermessung und -verringerung für Massenspeichersystem Download PDF

Info

Publication number
EP0370422A2
EP0370422A2 EP89121404A EP89121404A EP0370422A2 EP 0370422 A2 EP0370422 A2 EP 0370422A2 EP 89121404 A EP89121404 A EP 89121404A EP 89121404 A EP89121404 A EP 89121404A EP 0370422 A2 EP0370422 A2 EP 0370422A2
Authority
EP
European Patent Office
Prior art keywords
parameter
error
pulse
varying
read
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
EP89121404A
Other languages
English (en)
French (fr)
Other versions
EP0370422A3 (en
EP0370422B1 (de
Inventor
Larry J. Kendall
Andrew Palfreyman
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
National Semiconductor Corp
Original Assignee
National Semiconductor Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by National Semiconductor Corp filed Critical National Semiconductor Corp
Publication of EP0370422A2 publication Critical patent/EP0370422A2/de
Publication of EP0370422A3 publication Critical patent/EP0370422A3/en
Application granted granted Critical
Publication of EP0370422B1 publication Critical patent/EP0370422B1/de
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11C—STATIC STORES
    • G11C29/00—Checking stores for correct operation ; Subsequent repair; Testing stores during standby or offline operation
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
    • G11B20/00—Signal processing not specific to the method of recording or reproducing; Circuits therefor
    • G11B20/10—Digital recording or reproducing
    • G11B20/18—Error detection or correction; Testing, e.g. of drop-outs
    • G11B20/1816—Testing
    • G11B20/182—Testing using test patterns
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
    • G11B20/00—Signal processing not specific to the method of recording or reproducing; Circuits therefor
    • G11B20/10—Digital recording or reproducing
    • G11B20/10009—Improvement or modification of read or write signals
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
    • G11B20/00—Signal processing not specific to the method of recording or reproducing; Circuits therefor
    • G11B20/10—Digital recording or reproducing
    • G11B20/14—Digital recording or reproducing using self-clocking codes
    • G11B20/1403—Digital recording or reproducing using self-clocking codes characterised by the use of two levels
    • G11B20/1423—Code representation depending on subsequent bits, e.g. delay modulation, double density code, Miller code
    • G11B20/1426—Code representation depending on subsequent bits, e.g. delay modulation, double density code, Miller code conversion to or from block codes or representations thereof

Definitions

  • the invention relates to methods and apparatus for measuring and reducing the error rate of a mass storage device memory system. More particularly, the invention relates to methods and apparatus for measuring and reducing the error rate due to window centering, pulse pairing, and pattern sensitivity in a mass storage device memory system.
  • Figure 1 is a block diagram of a conventional magnetic disk drive system.
  • Disk 1 includes a number of concentric tracks such as track 2.
  • Magnetic read/write head 3 is separated by a small distance (typically about 0.00001 inches) from track 2, so that head 3 can read data from track 2 as disk 1 rotates relative to head 3.
  • the analog signal read by head 3 is amplified in read pre-amplifier 4 and supplied to pulse peak detector 5.
  • Pulse detector 5 differentiates the analog signal, and generates a digital signal whose pulses have leading edges corresponding to the zero-crossing points of the differentiated analog signal which correspond to the peaks and troughs of the incoming analog signal.
  • unit 5 may be any circuit which generates a digital signal whose leading edges correspond to the peaks and troughs of the incoming signal.
  • the digital signal emerging from unit 5 is then synchronized with a phase-locked VCO clock signal in synchronizer 7 to eliminate jitter, and is supplied to decoder 8 where it is converted to serial NRZ format suitable for processing in controller 9.
  • NRZ digital data to be written on disk 1 by head 3 is encoded into a desired format such as the well-known 2,7 RLL (Run Length Limited) code format in encoder unit 10.
  • 2,7 RLL code each bit representing a "one" is preceded and followed by a minimum of two zeros and a maximum of seven zeros.
  • the encoded data then undergoes precompensation in circuit 6 and is supplied to write driver 11, which generates appropriate signals to cause drive head 3 to write the data on disk 1.
  • Selection element 12 performs the function of selecting one of the read/write heads (in an embodiment including two or more independently selectable heads), selecting either a Read or Write mode, and (in the Write mode) supplying the proper write current to the head selected to write data onto disk 1.
  • Precompensation circuit 6 compensates for bit shift due to the tendency of closely spaced pulses to "repel" one another. This tendency causes two pulses written with minimum separation to be read back with more than one half the separation of two bits written with twice the minimum separation. Circuit 6 detects patterns (in the digital signal emerging from encoder 10) likely to exhibit such pattern-sensitive bit shifting. Then, while delaying the bits not selected by a nominal amount, circuit 6 delays the selected bits by more than a nominal amount (i.e., those selected bits likely to be "Early") or by less than a nominal amount (i.e., those selected bits likely to be "Late").
  • the above mentioned pattern sensitivity phenomenon (the repulsion between adjacent bits) is a significant source of error in reading from and writing onto a mass storage device (such as a disk or tape drive memory system).
  • the phenomenon is most pronounced at highest recorded bit densities, and accordingly varies in magnitude from track to track on a disk, having maximum magnitude for the track having minimum radius.
  • the phenomenon also varies with other parameters such as head variations, media variations, and head to media spacing.
  • window centering error there are two other significant sources of error in reading from and writing onto a mass storage device, namely window centering error and pulse pairing error. It is possible to define a parameter representing the contribution of each of pattern sensitivity error, window centering error, and pulse pairing error to the overall system error rate.
  • Window centering error is most readily characterized by the averaged time difference between the leading edge of each data pulse entering synchronizer 7 and the leading edge of each phase-locked VCO clock pulse.
  • the leading edge of the data pulses will ideally be positioned in time at the exact center of a time “window” defined by adjacent clock pulse leading edges.
  • the data will be strobed out (as "Sync Data) by the "Clock” leading edge following the "Data” leading edge.
  • Positioning the average “Data” leading edge at the exact center of the "window” allows any specific data pulse to be shifted forward or back by up to one half the window period and still be strobed out by the proper (next) clock pulse.
  • Increasing the average time separation between the clock edges and the data pulse centers will decrease the overall system error rate.
  • the "window center” can be adjusted by varying the relative delays of the Data pulse leading edges and the clock pulse leading edges.
  • Window centering error is purely a function of the accuracy of the Data and Clock relative delays in the synchronizer, and accordingly the operation of window centering need only be performed once for each drive employing multiple heads (not once for each read/write head).
  • Synchronizer 7 of the Fig. 1 system includes delay circuit 7a for shifting the phase of the phase-locked VCO clock pulses generated therein relative to the received Data pulse leading edges. It is conventional for the phase shift produced by delay circuit 7a to be variable in response to control signals supplied from controller 9 on line C1. Thus, the contribution of window centering error to the overall system error rate may be varied by such control signals on line C1.
  • Pulse pairing error results from the asymmetry between positive and negative flux transitions, which may be induced anywhere in the write-channel and read-­channel circuitry.
  • Figure 2 shows one way in which pulse pairing may arise.
  • Signal (a) has form typical of the amplified output of a magnetic disk drive read/write head, and includes positive polarity pulses 20, 22, and 24 alternating with negative polarity pulses 21, 23, and 25.
  • Magnets 13-17 comprise part of a sector of track 2 of disk 1 in Fig.1. Each track of disk 1 includes a plurality of independent sectors.
  • Signal (a) after it is differentiated in one of the circuit elements has the waveform of signal (b).
  • the zero-crossings of signal (b) correspond to peaks or troughs of signal (a).
  • Signal (d) represents the ideal output of a circuit for generating a train of pulses whose leading edges align with the zero-crossings of signal (b). In practice, the actual output of such a circuit will likely resemble signal (c), which exhibits pulse pairing.
  • a circuit intended to generate a stream of digital pulses from signal (b) will produce pulses whose leading edges align, not with zero-crossings of signal portions 30-35, but instead with points of signal (b) having a small amplitude offset as shown in Fig. 2.
  • the pulses of signal (c) are paired, with alternating long spacing T1 and short spacing T2 between adjacent pulses.
  • each pair of adjacent pulses associated with a positive polarity peak followed by a negative polarity peak i.e., pair of pulses 41 and 42
  • T1 the separation between adjacent pulses
  • each pair of adjacent pulses associated with a negative polarity peak followed by a positive polarity peak transition i.e., pair of pulses 42 and 43
  • T2 is less than T.
  • the error rate of a system is the number of bits processed incorrectly by the system divided by the number of bits processed correctly by the system.
  • conventional error tests are very time-consuming, commonly requiring on the order of several hours. Such conventional tests have been time-­consuming because typical error rates for disk drive systems are extremely low (for example 10 ⁇ 10), so that statistically significant accumulations of errors requires that many bits of data be read (at typical data rates in the range 5-20 Megabits/sec).
  • a practical, inexpensive method and means for rapidly and independently measuring and varying the error rate for each read/write head (or each storage medium track) has not been developed. Nor has a practical means been developed for independently varying the window centering, pulse pairing, and pattern sensitivity error parameters of a mass storage device memory system.
  • the inventive method is a rapid error measurement and reduction technique, including the steps of inducing a high mass storage device memory system error rate to determine independently the optimal pattern sensitivity, pulse pairing, and window centering parameters for the system.
  • the system is tuned by setting one or more of these parameters to its optimal value, so that the contribution of each optimized parameter to the system's overall error rate is minimized.
  • the inventive system includes means for performing the inventive method, and preferably includes computer-controllable means for independently adjusting the pattern sensitivity and pulse pairing error parameters associated with each read/write head of the system, and for adjusting the window centering error parameter associated with the system.
  • the inventive method exploits the mutual orthogonality (with respect to overall system error rate) of the pattern sensitivity, pulse pairing, and window centering error parameters in the following manner.
  • a high system error rate is induced by appropriately setting a second of the parameters.
  • the first parameter is varied around its nominal value until the value which produces a minimum error rate is found.
  • the process is repeated to determine the optimal value of the second parameter by appropriately setting the first (or third) parameter to induce a high overall system error rate and varying the second parameter to locate the value which produces a minimum error rate.
  • the process is repeated to determine the optimal value of the third parameter by setting the first (or second) parameter to induce a high system error rate and varying the third parameter for minimum error rate.
  • the three error parameters are "orthogonal" in the sense that the optimal value of any one of the three parameters does not depend on the values of the other two parameters.
  • a preferred embodiment of the inventive system includes circuitry for measuring the pulse pairing error parameter and for compensating for pulse pairing error by separately delaying positive and negative polarity signals read from the mass storage device, and adjusting these delays through separate, electronically adjustable delay lines or through a single rapidly adjustable delay line.
  • the window centering, pulse pairing, and pattern sensitivity error parameters of a mass storage device memory system are mutually orthogonal with respect to the system error rate, subject to the following qualification.
  • the window centering parameter (WC) and the pulse pairing parameter (PP) are always mutually orthogonal.
  • WC and the pattern sensitivity parameter (PC) will be mutually orthogonal only for certain test patterns (i.e., for certain sequences of bits read from or to be written onto a memory within the mass storage device).
  • PP and PC the pulse pairing and pattern sensitivity parameters, will be mutually orthogonal only for certain test patterns. The conditions under which WC and PC, and PC and PP will be orthogonal will be explained next.
  • signal (B) of Fig. 3 exhibits a randomized PP error phase with respect to the PC error phase (i.e., for pulse pair (2,3), the PP shift tends to reinforce the PC shift while for pulse pair (5,6), the PP shift tends to cancel the PC shift). Whatever deviations are made due to PP on pulse pair (2,3) are cancelled exactly on pulse pair (5,6). This will not have the effect of producing a net zero error rate due to PP and PC. Instead, the error rate will be the algebraic mean of the unsigned sum of the two error rates associated with the two different pairs of PP and PC parameters. Thus, for signal (B), the PC value found to produce the lowest error rate will not depend on the value of PP used to induce the high system error rate (to achieve short error rate measurement duration).
  • a "locked" test pattern such as signal (A) of Figure 3 should not be employed, and rather, a test pattern such as signal (B) of Figure 3 (i.e., an "unlocked” pattern exhibiting randomized PP phase relative to PC phase) should be employed.
  • the preferred "unlocked" test pattern may be precisely characterized with reference to the following model.
  • Signal (B) in Figure 3 is such a preferred test pattern.
  • the mapping of signal (B) in Figure 3 to NRZ will depend on the RLL code chosen (i.e., a 1,7 code, a 2,7, code, or some other code) and so is not specified herein.
  • RLL code chosen (i.e., a 1,7 code, a 2,7, code, or some other code) and so is not specified herein.
  • other, more complex "unlocked" test patterns satisfying the specified criteria may be used instead of signal (B) of Figure 3, especially if difficulties exist in encoding the signal (B) pattern from NRZ.
  • the "PC/PP" steps of the inventive method may be optimized for different values of the fixed error parameter (i.e., for different bit-pair spacings, when PC is held fixed).
  • the fixed error parameter i.e., for different bit-pair spacings, when PC is held fixed.
  • each test pattern employed may have a uniform bit-pair spacing, so that the bit-pair spacing remains substantially constant during each performance of the method.
  • one test pattern for each read/write head may be fixed on a track of a disk (or, more generally, on a storage location in a mass storage device memory system) so that each test pattern may be read by a different read/write head during performance of the test.
  • the method will preferably be performed independently for each head, using separate hardware for each head.
  • the preferred "unlocked" test pattern also satisfies the criterion that the quantity (d1- d2) is maximized.
  • a "locked” test pattern such as signal (A) of Figure 3, or an “unlocked” test pattern such as signal (B) of Figure 3 should be employed.
  • FIG. 4 is a block diagram of a magnetic disk drive system with such capability.
  • the Fig. 4 system includes many of the same elements of the conventional Fig. 1 system, but differs from the Fig. 1 system in the following respects.
  • Circuit 6 of Fig. 1 is replaced by precompensation and pulse pairing compensation circuit 6′ in Fig. 4.
  • Circuit 6′ will be discussed in detail below with reference to Figures 5 through 16.
  • Circuit 6′ is capable of independently varying the system's PP and PC error parameters in response to control signals supplied from controller 9 on line C2.
  • An artificially large overall system error rate may be induced by intentionally changing one of the error parameters away from its optimal value. Then, the optimal value of another of the error parameters may be determined.
  • the overall system error rate may be minimized by sequentially determining the optimal values of the three error parameters and then setting all three of them to their optimal values.
  • the embodiment of the inventive system shown in Fig. 4 has a second read/write head 103 identical to head 3, which may be positioned to read data from and write data onto a track of disk 1 other than track 2, or a track of a disk (not shown) other than disk 1.
  • Controller 9 is capable of switching circuit 12 to accept data from either head 3 or head 103.
  • controller 9 is capable of switching circuit 12 to supply data either to head 3 or head 103.
  • each of heads 3 and 103 will thus be associated with a different PP parameter and with a different PC parameter, which may be independently varied by circuit 6′ when head 3 or head 103 are selected respectively.
  • systems embodying the invention may alternatively include less than two or more than two read/write heads. Where more than two heads are included, it is contemplated that all will preferably be controlled by a common controller, and will share the same pulse detector, precompensation and pulse pairing correction circuitry, synchronizer, decoder and encoder circuitry.
  • a high system error rate is induced by appropriately varying a first one of the parameters WC, PP, and PC.
  • the system error rate is then measured for each of a plurality of different values of a second one of the parameters WC, PP, and PC.
  • controller 9 This is done by generating appropriate control signals in controller 9 to cause the read/write head repeatedly to read the test pattern on the mass storage medium (which may be a magnetic disk as shown in Fig. 4), while the controller sends control signals at appropriate intervals to change the value of the "second" parameter, and while the error rate is computed in the controller (or in a computer interfaced therewith) during each interval in which the second parameter remains unchanged.
  • the optimal value of this second parameter is identified as the value resulting in the lowest overall system error rate.
  • the process is repeated to determine the optimal value of the first parameter by appropriately varying the second (or third) parameter to induce a high overall system error rate, then measuring system error rate for each of a plurality of different values of the first parameter, and identifying the "optimal" value of the first parameter as the one resulting in the lowest overall system error rate.
  • the process is repeated a third time to determine the optimal value of the third parameter by varying the first (or second) parameter to induce a high system error rate, and then measuring the overall system error rate for each of a plurality of values of the third parameter.
  • the system controller sets each of the parameters PP, PC, and WC to its optimal value.
  • the parameter PP is first held fixed while the parameter WC is varied (to determine the optimum value of parameter WC). Then, the parameter WC is held fixed while PP is varied. Finally, the parameter WC (or PP) is held fixed while PS is varied. As explained above, special test patterns must be employed when WC is held fixed and PC varied (or when PC is held fixed and WC varied), or when PP is held fixed and PC is varied (or when PC is held fixed and PP is varied).
  • a desirable feature of the invention is that artificially high error rates may be induced by appropriately choosing the value of the parameter held fixed in each step of the inventive method.
  • the inventive system need not be designed and operated so as to measure extremely low error rates with great accuracy. Instead the system need only be capable of measuring relatively high error rates.
  • Figure 5 is a block diagram of a preferred embodiment of circuit 6′ of Fig. 4. Most components of the Fig. 5 circuit have a dual role in the sense that they are used to perform precompensation in the system's "write” mode and to perform pulse pairing compensation in the system's "read” mode.
  • digital data in the 2,7 RLL format (the signal labeled "DATA 2 7") is received from encoder 10.
  • the Fig. 5 circuit "precompensates” the data by recognizing sensitive bit patterns likely to exhibit bit repulsion and then for each sensitive pattern, delaying the early bits relative to the nominal bit delay, and writing the late bits early relative to the nominal bits.
  • the Fig. 5 In the system's "read” mode, the Fig.
  • circuit performs pulse pairing compensation by recognizing the polarity of each incoming bit (i.e., the circuit recognizes whether the associated input pulse to pulse peak detector 5 shown in Figure 4 had positive or negative polarity) and delaying each incoming bit by a preselected amount depending on its polarity.
  • the Fig.5 circuit receives control signals from controller 9 that determine the circuit's operating mode ("write” or "read”) and the amount by which each bit category be delayed. Controlling the amount of delay for each bit category is equivalent to controlling the PP and PC error parameters for the system.
  • Circuit 100 accordingly outputs a signal on the appropriate one of lines PE (pulse early), PZ (pulse nominal), or PL (pulse late).
  • "VTH SELECT" circuit 101 receives such signal and outputs a corresponding one of three voltage values on line VT to ECOMP circuit 107.
  • the three voltage values represent late delay and early delay variations from a nominal delay, in addition to the nominal delay itself.
  • the three voltage values are variable in response to signals received from controller 9 on the mode bus identified as "MD BUS.”
  • Circuit 100 also outputs the bit to be precompensated and supplies it on line PRECOMP to circuit 104, which is a CMOS-to-ECL translator.
  • the signal emerging from circuit 104 is then selected by DECL MUX circuit 106, and supplied to RC circuit 113.
  • Circuit 113 upon receipt of the PRECOMP pulse from circuit 106, allows its output voltage to ramp down at a rate determined by an external resistor and capacitor and applies this ramp to the input of ECOMP circuit 107. When the ramp equals the selected voltage value received on line VT, circuit 107 will switch its outputs.
  • each bit in turn is delayed by a nominal amount, or by a pre-­selected EARLY amount or a pre-selected LATE amount.
  • the EARLY and LATE amounts are pre-selected by signals (to be discussed below with reference to Fig. 7) supplied from controller 9 via the MD BUS.
  • Fig. 5 In the "read" mode of the Fig. 5 circuit, data is input from pulse detector 5 (shown in FIg. 4) on line "INPUT 2 7," and is supplied to bypass multiplexer 112, to EARLY LATE circuit 100, and to translator circuit 103.
  • the function of translator 103 is to translate the voltage level to the ECL level required by flip-flop 105.
  • the signal output from circuit 103 is supplied through flip-flop 105, DECLMUX circuit 106, and RC circuit 113 to ECOMP circuit 107.
  • the function of circuit 105 is to terminate the pulse started by the leading edge of the signal "INPUT 2 7.”
  • the function of circuit 106 is to select WRT pulses from the other pulses emerging from circuits 104 and 105.
  • a polarity signal (“PULSE POL”) from pulse detector 5 is also supplied to EARLY LATE circuit 100.
  • circuit 100 determines the polarity associated with each incoming bit and accordingly outputs a signal on either line PE or line PZ to identify the bit as an early or late bit.
  • the PE or PZ signal is supplied to VTH SELECT circuit 101, where a pre-selected voltage value is chosen (from a menu determined by control signals received from controller 9 on the MD BUS).
  • the chosen voltage value (which corresponds to a pre-selected delay) is supplied to ECOMP circuit 107 as in the write mode.
  • RC circuit 113 again provides a voltage ramp to comparator 107, and comparator 107 changes state when the ramp decreases to the VT voltage value chosen in circuit 101.
  • the output pulse (labeled “RC 2 7”) from comparator 107 goes through translator 108 and multiplexer 112 to synchronizer 7′ via line "C IN 2 7.”
  • Output pulse “RC 2 7” is also supplied to EARLY LATE circuit 100 to update the polarity (for the next bit) by simply selecting the opposite polarity to that of the present bit.
  • the output of circuit 107 is also supplied to buffer circuit 109.
  • the function of circuit 109 is to buffer the signal from circuit 107.
  • the output of circuit 109 is supplied to circuit 105 for the purpose of terminating the pulse started by the "INPUT 2 7" signal.
  • EARLY LATE circuit 100 is shown in Figure 6.
  • the circuitry within block 200 selects the pulse pairing compensation mode of operation for circuit 100 and selects the proper polarity for compensation.
  • Signal MD3 arriving on the MD BUS instructs circuit 100 whether the positive polarity or the negative polarity incoming bits are to be delayed by a less than nominal amount.
  • the nominal delay amount, and the "less than nominal” and “more than nominal” delay amounts are also set by signals, to be discussed below with reference to Fig. 6, arriving on the MD BUS.
  • the output of flip-flop 201 is the signal "PULSE LOCK,” which is clocked into flip-flop 202 by the delayed "RC 2 7" signal, causing the output of circuit 202, signal “FDLY ER,” to change after each delayed pulse.
  • signal WRT GATE When in the "read” mode, signal WRT GATE will be a logical “not true” and thus will deselect transmission gates 203 and 204 and will select transmission gates 205 and 206. In this mode, PL will accordingly always be "high” (i.e., equal to +VCC) and PE and PZ will alternate according to FDLY ER via gate 205 and inverting NAND gate 207. This will alternately select EARLY and Nominal compensation for alternating ones of the bits comprising signal INPUT 2 7.
  • transmission gates 203 and 204 are activated (and gates 205 and 206 are deactivated) and the pulse pairing circuitry deactivated by signal WRT GATE 1.
  • the precompensation circuitry within block 210 accepts data signal DATA 2 7, and clocks this signal into seven-bit shift register 212 with delayed clock signal PRECOMP CLK. Clock signals "CLK G" to be discussed below, and PRECOMP CLK arrive on the CLK BUS from encoder 10. PRECOMP CLK is then delayed in buffer 211.
  • the state of shift register outputs A through G is monitored by three-input NAND gates 213 and 214 and flip-flops 215, 216 and 217.
  • Flip-flop 217 outputs a "one” whenever a "one” (the bit to be compensated) appears at position D. This "one" at the output of circuit 217 enables latches FDLY EH and FDLY LH.
  • the relationships of signals PDLY E, PDLY L, and PULSE CLR to the state of shift register 212's outputs is shown in the following table: A B C D E F G PDLY E PDLY L PULSE CLR x x x 0 x x 0 0 0 0 0 1 0 0 0 0 0 0 1 1 0 0 1 0 0 1 0 0 0 0 0 1 0 0 1 0 0 0 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1
  • PDLY L will be a "one” if bit A is a “one” and bit G is a zero when bit D is a “one”; otherwise PDLY L will be a "zero".
  • Pulse CLR is zero whenever bit D is zero allowing the latches to hold the set values until the next "one,” which will enable them to change.
  • Flip-flop 219 provides a pulse starting at the clock pulse after the clock pulse that positioned the "one" in position D.
  • the pulse output by circuit 219 has duration equal to two clock periods due to the presence of NOR gate 220 between gates E and F and circuit 219.
  • the PRECOMP signal emerging from circuit 219 is supplied to translator 104 as shown in Fig. 5.
  • the precompensation scheme of Figs 5 and 6 thus compensates a bit when the bit is preceded by the sequence "zero zero one” and followed by three or more "zeroes,” and when the bit is followed by the sequence "one zero zero” and preceded by three or more “zeroes.” No compensation is performed when there are three or more "zeroes” preceding and following the bit, and no compensation is performed when there are two “zeroes” and a "one” both preceding and following the bit.
  • the rationale behind this scheme is that a first bit will be repelled by the near presence of an adjacent second bit unless there are bits in close proximity to the first bit on both sides of the first bit. For a bit sequence "zero zero zero one zero zero one zero zero zero zero zero", the first "one” will be written late (closer to the middle "one”), the middle "one will not be compensated, and the third "one” will be written early.
  • Figure 7 shows VTH SELECT circuit 101 of Fig. 4.
  • Circuit 101 receives signals MD 0, MD 1, and MD 2 (for pulse pairing compensation) and signals MD 5, MD 6, and MD 7 (for precompensation) on the MD BUS.
  • Circuit 101 also receives signal WRT GATE 1, and supplies WRT GATE 1 to transmission gate pairs 300, 301, and 302.
  • circuits 300, 301, and 302 In the "write” mode, circuits 300, 301, and 302 will supply signals MD 5, MD 6, and MD 7 to decoder circuit 303.
  • circuits 300, 301, and 302 will supply signals MD 0, MD 1, and MD 2 to circuit 303.
  • circuit 303 The input signals to circuit 303 are decoded to select one of signals P0 through P6, or "BYPASS.” If the output of circuit 303 is selected to be BYPASS, then the Fig. 5 circuit will not perform precompensation on incoming signal DATA 2 7. If the output of circuit 303 is P0, the output of circuit 101 (VTH) is supplied from the center of resistor string 304 via transmission gates 305 and 306 and output emitter follower 307. This path is also selected whenever signal PZ from circuit 100 is in a "true" state, regardless of the MD bit status.
  • MD bits (MD 0 through MD 7) select other than BYPASS or PO (i.e., if they select one of signals PI through P6)
  • an EARLY value is selected from the upper bank of resistors in string 304
  • a LATE value is selected from the lower bank of resistors in string 304, by selecting the appropriate ones of transmission gates 308.
  • These EARLY and LATE voltages are supplied to transmission gates 309 and 310, respectively, which are selected by PE and PL when PO is in a "not true" state.
  • the logic of EARLY LATE circuit 100 ensures that signals PE, PZ, and PL are mutually exclusive.
  • Fig. 8 is a preferred embodiment of circuit 103 of Fig. 5.
  • Fig. 9 is a preferred embodiment of circuit 104 of Fig. 5.
  • Fig. 10 is a preferred embodiment of circuit 105 of Fig. 5.
  • Fig. 11 is a preferred embodiment of circuit 106 of Fig. 5.
  • Fig. 12 is a preferred embodiment of circuit 107 of Fig. 5.
  • Fig. 13 is a preferred embodiment of circuit 108 of Fig. 5. It is important that resistors M1 in circuit 108 match (i.e., they must have identical dimensions and orientation). Similarly, resistors M2 in circuit 108 must match.
  • Fig. 14 is a preferred embodiment of circuit 109 of Fig. 5.
  • Fig. 15 is a preferred embodiment of identical circuits 110, 111, and 112 of Fig. 5.
  • Fig. 16 is a preferred embodiment of circuit 113 of Fig. 5. It is important that the collector of transistor 9′ of Fig. 16 be connected directly to the VCC pad to prevent switching currents from upsetting other internal circuitry.
  • pulse pairing may be measured, and pulse pairing compensation performed, using circuitry separate from the precompensation circuitry. For example, either operation may be accomplished in pulse detector 5 of Fig. 4.
  • Figure 17 is an example of such a circuit that may be included in pulse detector 5 for measuring the magnitude of pulse pairing error in a stream of data read from a mass storage device memory system.
  • a "Polarity" signal (indicative of the polarity associated with each incoming data pulse) is supplied to the clock input of flip-flop 400, and a "Preamble" signal indicative of reading a preamble field of a fixed frequency or spacing is supplied to the data input of flip-flop 400.
  • the output of flip-flop 400, and the incoming data stream (the "Data" signal) are supplied to flip-flop 401.
  • the output of 401 goes to multiplexer 402.
  • the output of 402 is the Reference “IF” frequency until the first Data pulse, related to a "positive” Read pulse, following the assertion of PREAMBLE (at which time the output of 402 becomes the Data stream).
  • the output of flip-flop 401 also activates one-shot 404.
  • the output of circuit 404 clears flip-flop 403 thereby ensuring the same initial conditions for each reading.
  • the output of flip-flop 401 also goes to one input terminal of AND gate 405.
  • the output of 405 will be true at the delayed PREAMBLE from flip-flop 401 and will go false at the end of PREAMBLE.
  • the output of AND gate 405 will open switch 410 during this shortened PREAMBLE and at the end of this shortened PREAMBLE will activate one-shot 406 which will open switch 411 for the period of time set by one-shot 406.
  • the output of one-­shot 406 will establish a "READ OUTPUT" time window.
  • Flip-flop 403 when driven by "Reference + 2" during NON-­PREAMBLE time, will alternately assert Q and Q for equal times deviating only by any unequal delays of itself (flip-flop 403).
  • Resistor R1 will charge capacitor C1, ideally to V Low + (V High - v Low )/2 assuming a perfect match of Q time and Q .
  • Resistor R2 will charge capacitor C2 to the same value related to Q .
  • the difference of the voltages on capacitors C1 and C2 will be amplified by difference amplifier 407 and applied to capacitors C3 and C4 via resistors R3 and R4 and switches 410 and 412. The voltages on capacitors C3 and C4 will be identical at this time.
  • the Data will be the recovered signal from a constant frequency pattern approximately equal to Reference + 2 and will include any inherent pulse pairing.
  • Flip-flop 403's outputs will change their time relationship to include the pulse pairing which will result in different charges on C1 and C2. This difference, again, is amplified by 407.
  • C3 will store the initial balanced condition while C4 will charge to the new value containing the pulse-pairing information.
  • Differential amplifier 408 will amplify this difference and apply it to buffer amplifier 409 via switch 411 and resistor R5.
  • the input to amplifier 409 is filtered by capacitor C5.
  • Output D from amplifier 409 will then represent only the pulse pairing (T1-T2) of the Data signal.
  • switch 411 will open and the output signal D may then be read by conventional means.
  • the output of multiplexer 402 is supplied to the clock input of flip-flop 403.
  • the voltage at point D is thus proportional to the quantity T1- T2, where T1 is the time delay between a positive polarity data pulse and the nearest subsequent negative polarity data pulse, and T2 is the time delay between a negative polarity data pulse and the dearest subsequent negative polarity data pulse.
  • voltage signal D is an error signal indicative of the magnitude of pulse pairing error.
  • Signal D may be used to generate control signals for a circuit such as that shown in Figure 18 to vary the difference, T1-T2, between the two delays as desired, for example to compensate for the effect of pulse pairing.
  • Signal D may alternatively be read by an analog to digital converter and subsequently held in logic, such as in controller 9, where it may be used to generate a PP error parameter, and to generate control signals (such as MD 0, MD 1, and MD 2 discussed above) for compensating for the pulse pairing error.
  • Figure 18 is a circuit for performing pulse pairing error compensation on a stream of data read from a mass storage device memory system.
  • the data is supplied to an input of each of AND gates 501 and 502, and to polarity detector circuit 500.
  • Circuit 500 outputs a signal indicative of the polarity associated with each pulse of the incoming data stream to the other input of each of AND gates 501 and 502.
  • Variable delay lines 504 and 503 are independently controllable in response to control signals E and F, respectively.
  • the compensated data emerges from NOR gate 505.
  • Each of delay lines 503 and 504 may be a silicon delay line such as employed in National Semiconductor Corporation's "DP845" integrated circuit.
  • the output signal D from the Fig. 17 circuit may be used to generate control signals E and F as described in the previous paragraph.

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Signal Processing For Digital Recording And Reproducing (AREA)
  • Digital Magnetic Recording (AREA)
  • Test And Diagnosis Of Digital Computers (AREA)
EP89121404A 1988-11-23 1989-11-18 Verfahren und Gerät zur Fehlermessung und -verringerung für Massenspeichersystem Expired - Lifetime EP0370422B1 (de)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US276209 1988-11-23
US07/276,209 US4933939A (en) 1988-11-23 1988-11-23 Method and apparatus for error measurement and reduction in a mass storage device memory system

Publications (3)

Publication Number Publication Date
EP0370422A2 true EP0370422A2 (de) 1990-05-30
EP0370422A3 EP0370422A3 (en) 1990-09-19
EP0370422B1 EP0370422B1 (de) 1995-02-01

Family

ID=23055655

Family Applications (1)

Application Number Title Priority Date Filing Date
EP89121404A Expired - Lifetime EP0370422B1 (de) 1988-11-23 1989-11-18 Verfahren und Gerät zur Fehlermessung und -verringerung für Massenspeichersystem

Country Status (5)

Country Link
US (1) US4933939A (de)
EP (1) EP0370422B1 (de)
JP (1) JPH02210671A (de)
KR (1) KR0138779B1 (de)
DE (1) DE68920960T2 (de)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0488276A3 (en) * 1990-11-30 1993-01-20 Hitachi, Ltd. Diagnostic system of magnetic recording channel and magnetic disk drive apparatus
EP0571905A3 (de) * 1992-05-25 1995-06-07 Sony Corp Modulationsverfahren.
EP0581717A3 (en) * 1992-07-30 1995-12-20 Ibm Error measurement circuit
CN114356229A (zh) * 2021-12-22 2022-04-15 合肥康芯威存储技术有限公司 一种数据存储设备的参数优化方法及其优化系统

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH02239474A (ja) * 1989-02-03 1990-09-21 Deutsche Thomson Brandt Gmbh オーデイオまたはビデオ信号のデイジタル記録および再生用の装置
US5163051A (en) * 1990-02-02 1992-11-10 Telecom Analysis Systems Inc. Paired bit error rate tester
US5422890A (en) * 1991-11-19 1995-06-06 Compaq Computer Corporation Method for dynamically measuring computer disk error rates
US5408367A (en) * 1993-09-20 1995-04-18 Integral Peripherals, Inc. Method of optimizing operation of disk drive
US5574855A (en) * 1995-05-15 1996-11-12 Emc Corporation Method and apparatus for testing raid systems
US5787086A (en) * 1995-07-19 1998-07-28 Fujitsu Network Communications, Inc. Method and apparatus for emulating a circuit connection in a cell based communications network
US6662329B1 (en) 2000-03-23 2003-12-09 International Business Machines Corporation Processing errors in MPEG data as it is sent to a fixed storage device
US8214589B2 (en) * 2009-03-13 2012-07-03 Seagate Technology Llc Data storage system redundancy scheme verification
JP2019046523A (ja) 2017-09-04 2019-03-22 株式会社東芝 磁気ディスク装置及びライト処理方法
CN112767977B (zh) * 2020-12-31 2023-09-26 深圳市紫光同创电子有限公司 读写窗口校准电路及方法、存储器、fpga芯片

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4024478A (en) * 1975-10-17 1977-05-17 General Electric Company Printed broadband A. C. grounded microwave terminations
US4480274A (en) * 1982-04-16 1984-10-30 Ampex Corporation Record verify and optimize system
US4841526A (en) * 1984-05-25 1989-06-20 Wilson Jon C Data communications system
US4821125A (en) * 1986-06-11 1989-04-11 International Business Machines Corporation Compensation of write current adn delta-V for recording component and radial position
JPS62298255A (ja) * 1986-06-18 1987-12-25 Fujitsu Ltd 識別装置
US4829545A (en) * 1986-08-25 1989-05-09 Guzik Technical Enterprises, Inc. Method and apparatus for data window centering ina multifrequency data separator

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0488276A3 (en) * 1990-11-30 1993-01-20 Hitachi, Ltd. Diagnostic system of magnetic recording channel and magnetic disk drive apparatus
US5459620A (en) * 1990-11-30 1995-10-17 Hitachi, Ltd. Diagnostic system of magnetic recording channel and magnetic disk drive apparatus
EP0571905A3 (de) * 1992-05-25 1995-06-07 Sony Corp Modulationsverfahren.
CN1061159C (zh) * 1992-05-25 2001-01-24 索尼公司 调制方法
EP0581717A3 (en) * 1992-07-30 1995-12-20 Ibm Error measurement circuit
CN114356229A (zh) * 2021-12-22 2022-04-15 合肥康芯威存储技术有限公司 一种数据存储设备的参数优化方法及其优化系统
CN114356229B (zh) * 2021-12-22 2023-09-22 合肥康芯威存储技术有限公司 一种数据存储设备的参数优化方法及其优化系统

Also Published As

Publication number Publication date
DE68920960T2 (de) 1995-09-28
DE68920960D1 (de) 1995-03-16
US4933939A (en) 1990-06-12
EP0370422A3 (en) 1990-09-19
JPH02210671A (ja) 1990-08-22
KR0138779B1 (ko) 1998-06-01
KR900008529A (ko) 1990-06-04
EP0370422B1 (de) 1995-02-01

Similar Documents

Publication Publication Date Title
US5517146A (en) Write precompensation circuit for compensating nonlinear magnetic effects in a recording channel
US5420893A (en) Asynchronous data channel for information storage subsystem
EP0370422B1 (de) Verfahren und Gerät zur Fehlermessung und -verringerung für Massenspeichersystem
US5917668A (en) Synchronous read channel employing a frequency synthesizer for locking a timing recovery phase-lock loop to a reference frequency
EP0706173B1 (de) Optimierung der Schreibprekompensation bei einem PRML-Kanal
JP2868707B2 (ja) ディスク状記録担体を利用したデータ記録装置の読出チャンネル最適化方法
US6337778B1 (en) Disk drive employing vector addition of primary phase write clock signals for generating secondary phase write clock signal
US6208477B1 (en) Hard disk drive having a built-in self-test for measuring non-linear signal distortion
US5995305A (en) Disk drive using off-track margin to obtain optimal performance parameters
EP0701255A2 (de) Nullphasenanlaufoptimierung nach dem Verfahren des Mittelwertquadratfehlers in einem PRML-Aufzeichnungskanal
US20020176188A1 (en) Offset cancellation of charge pump based phase detector
US5311376A (en) Information detecting system
JP3274490B2 (ja) コンピュータデータテープ駆動機構のための書込回路、パルス回路および書込等化パルスを調整するための方法
EP0809246B1 (de) Verbesserungen in Bezug auf Datenverarbeitung
US20020015247A1 (en) High resolution wide range write precompensation
EP0222132A2 (de) Trennschaltung für digitale Daten
US6438703B1 (en) Method and system for selectively varying signal delay in response to detection of a quiescent signal
US6404577B1 (en) Trellis encoding and decoding of track identification indicia
US5335118A (en) Disc drive with write precompensation using window/pulse timing measurements
EP1040474A2 (de) Adaptive und selektive pegelkonditionierung eines lesekanals in speichertechnologien
US6313962B1 (en) Combined read and write VCO for DASD PRML channels
JPH0877503A (ja) ピーク検出回路およびそれを使用した記録媒体再生装置
EP0959470A1 (de) Kalibrierung der Schreib-Vorentzerrung in einer Datenspeicheranordnung
JPS61204870A (ja) 磁気記録再生装置のインデツクスマ−ク検出方式

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

AK Designated contracting states

Kind code of ref document: A2

Designated state(s): DE FR GB IT NL

PUAL Search report despatched

Free format text: ORIGINAL CODE: 0009013

AK Designated contracting states

Kind code of ref document: A3

Designated state(s): DE FR GB IT NL

17P Request for examination filed

Effective date: 19910124

17Q First examination report despatched

Effective date: 19930426

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): DE FR GB IT NL

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: IT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRE;WARNING: LAPSES OF ITALIAN PATENTS WITH EFFECTIVE DATE BEFORE 2007 MAY HAVE OCCURRED AT ANY TIME BEFORE 2007. THE CORRECT EFFECTIVE DATE MAY BE DIFFERENT FROM THE ONE RECORDED.SCRIBED TIME-LIMIT

Effective date: 19950201

Ref country code: NL

Effective date: 19950201

Ref country code: FR

Effective date: 19950201

REF Corresponds to:

Ref document number: 68920960

Country of ref document: DE

Date of ref document: 19950316

EN Fr: translation not filed
NLV1 Nl: lapsed or annulled due to failure to fulfill the requirements of art. 29p and 29m of the patents act
PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: GB

Effective date: 19951118

PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

26N No opposition filed
GBPC Gb: european patent ceased through non-payment of renewal fee

Effective date: 19951118

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: DE

Payment date: 20060102

Year of fee payment: 17

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: DE

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20070601